Polish the Gold: Calibrating Intervention Effect Estimates for Small-Sample Randomized Controlled Trials